https://github.com/dfm/optimize.js

A tool for robust numerical optimization running right in your browser
https://github.com/dfm/optimize.js

Last synced: about 2 months ago
JSON representation

A tool for robust numerical optimization running right in your browser

• Host: GitHub
• URL: https://github.com/dfm/optimize.js
• Owner: dfm
• Created: 2012-07-09T20:58:52.000Z (about 12 years ago)
• Default Branch: main
• Last Pushed: 2020-06-12T18:24:25.000Z (over 4 years ago)
• Last Synced: 2024-07-01T21:15:04.875Z (3 months ago)
• Language: JavaScript
• Homepage: http://danfm.ca/optimize.js
• Size: 87.9 KB
• Stars: 20
• Watchers: 6
• Forks: 7
• Open Issues: 1
• Metadata Files:
  • Readme: README.markdown

Awesome Lists containing this project

README

        
optimize.js

===========

[This](http://en.wikipedia.org/wiki/Mathematical_optimization) kind of

optimization. Not [that](https://developers.google.com/closure/compiler/).

This was ported from the `scipy.optimize` and we seem to be about an order

of magnitude faster than the Python version but that hasn't been too well

tested.

Usage

-----

To optimize some dumb function, just run

```javascript

my_function = function (x) {

    return optimize.vector.dot(x, x);

};

xopt = optimize.fmin(my_function, [5.0, -3.4, 1.7, 16.3, 0.17]);

```

And this should say something like:

```

Converged in 349 iterations.

Function value = 3.954810202072493e-7

```

And `xopt` should end up being something like

```

[-0.0003993727670733724, -0.00027257793115893254, -0.0003811443958917447, 0.00012560283306980614, 0.00002523018782488204]

```

You can also include some options as follows:

```javascript

xopt = optimize.fmin(my_function, [5.0, -3.4, 1.7, 16.3, 0.17], {ftol: 1e-7, maxiter: 1000});

```

To Do

-----

* Write tests.

* Implement some better algorithms.

* Make some demos.